[R-sig-eco] forward selection of significant eigenvectors

[Stéphane Dray]

Yes ! Perfect.


Le 06/06/2019 à 17:17, Tim Richter-Heitmann a écrit :
> Hi Stephane,
>
> thank you for this very useful information.
>
> As for the global test, the tutorial of Benoit Gendreau-Berthiaume (as
> linked in my mail) does start the analysis with anova.cca on a RDA
> fitting all available dbMEM before doing variable selection.
>
> Does this satisfy the requirement of a global test?
>
> Cheers, Tim
>
> Am 06.06.2019 um 17:03 schrieb Stéphane Dray:
>> Hi,
>>
>> we introduced a new function mem.select in adespatial (that used
>> forward.sel)  based on the work by Bauman et al
>> (https://onlinelibrary.wiley.com/doi/abs/10.1111/ecog.03380)
>> Basically, we showed that you should use a global test before
>> performing forward selection (this step is suggested by Blanchet et al
>> but rarely performed  in ecological studies).
>>
>> Cheers
>>
>>
>> Le 06/06/2019 à 16:41, Tim Richter-Heitmann a écrit :
>>> Dear Irene,
>>>
>>> i found this tutorial very useful:
>>>
>>> https://sites.ualberta.ca/~ahamann/teaching/renr690/Lab9.pdf
>>>
>>> You basically do a forward selection (forward.sel() from packfor
>>> package) using the adjusted R-square (RsquareAdj() from vegan package).
>>>
>>> Hope this helps, Tim
>>>
>>> Am 06.06.2019 um 16:34 schrieb Irene Adamo:
>>>> Dear all,
>>>>
>>>> I am trying to run a variaion partitioning in using the varpart
>>>> function in
>>>> vegan. I would like to include the geographic coordinates as
>>>> explanatory
>>>> variables so I calculated the principle coordinates of neighbour
>>>> matrices.
>>>> Now I would like to perform a selection of the significant spatial
>>>> eigenvectors but I don`t know how to do it in R. Could anybody help me?
>>>>
>>>> data(mite.xy)
>>>> pcnm1 <- pcnm(dist(mite.xy))
>>>>
>>>> then I get the pcnm vectors and I would like to select only the
>>>> significant ones.
>>>>
>>>> Thanks a lot!
>>>>
>>>>      [[alternative HTML version deleted]]
>>>>
>>>> _______________________________________________
>>>> R-sig-ecology mailing list
>>>> R-sig-ecology using r-project.org
>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology


-- 
Stéphane DRAY (stephane.dray using univ-lyon1.fr)
Laboratoire BBE-CNRS-UMR-5558, Univ. C. Bernard - Lyon I
43, Bd du 11 Novembre 1918, 69622 Villeurbanne Cedex, France
http://pbil.univ-lyon1.fr/members/dray/
GdR EcoStat : https://sites.google.com/site/gdrecostat/